1. Field of the Invention
The present invention relates generally to aligners and patterning methods using a phase shift mask, and more particularly to an aligner and a patterning method by interference of light, using a phase shift mask for photoresist.
2. Description of the Background Art
High density integration and miniaturization in the field of semiconductor integrated circuit devices have remarkably advanced in recent years. Circuit patterns formed on semiconductor substrates (hereinafter referred to as wafers) have rapidly shrinked accordingly.
Among others, the photolithography technique has been widely recognized as a basic technique in patterning. This is why various developments and improvements have been made in this field. The shrinking of patterns seems to have no end, and there exists even a stronger demand for higher resolution of the patterns.
The photolithography technique is a patterning technique according to which a mask pattern (original picture) is transferred onto photoresist applied on a wafer, and the photoresist with the transferred pattern is used to pattern a film to be etched under the photoresist. When the pattern is transferred onto the photoresist, the photoresist is subjected to development, during which, with positive type photoresist, the exposed areas are removed, while with negative type photoresist, the unexposed areas are removed.
In general, the resolution limit R (nm) in the photolithography technique by means of demagnification exposure is represented as follows: EQU R=k.sub.1 .multidot..lambda./(NA)
where .lambda. is the wave length (nm) of light, NA is the numerical aperture of a lens, and k.sub.1 is a constant depending on the resist process.
As can be seen from the above expression, the resolution limit R may be reduced by reducing the values of k.sub.1 and .lambda. while increasing the value of NA, in other words, by reducing the constant k.sub.1 depending on the resist process and shortening the wavelength of exposure light and increasing the numerical aperture of the lens.
It is however technically difficult to improve light sources and lenses, and such shortened wave length and higher NA may reduce the depth of focus .delta. (.delta.=k.sub.2 .multidot..lambda./(NA).sup.2) of light, resulting in lower resolution instead.
This is why improvement of photomasks and shrink patterns rather than light sources or lenses has been studied. In recent years, the phase shift mask has attracted much attention as a photomask in the super-resolution technique for improving pattern resolution. The structure and principles of such masks as half tone type and Levenson-type will be described in comparison with usual masks by way of illustration. As for the mask pattern, an entire pattern will be described.
In FIG. 15, (a) and (b) are a top and a cross section of a usual photomask, respectively, and (c) and (d) show the electric field over the photomask in use and the optical intensity of an optical image on the wafer. Referring to FIG. 15, at (a) and (b), the usual photomask has a metal mask pattern 103 formed on a glass substrate 101. In such a photomask, the electric field over the mask is spacially pulse-modulated by metal mask pattern 103 as shown in FIG. 15 at (c).
Referring to FIG. 15 at (d), with a pattern reduced in size, exposure light transmitted through the photomask reaches around the unexposed region (region in which the transmission of exposure light is interrupted by metal mask pattern 103) on the wafer because of the diffraction of light. Therefore, the unexposed region of the wafer is also irradiated with light, resulting in a reduced contrast of light (difference in optical intensity between the exposed region and the unexposed region of the wafer). As a result, the resolution degrades, which makes it difficult to transfer fine patterns.
In FIG. 16, (a) and (b) show a top and a cross section of a half tone type phase shift mask, respectively, and (c) and (d) show the electric field over the mask and the optical intensity of an optical image on the wafer, respectively. Referring to FIG. 16 at (a) and (b), the half tone type phase shift mask is provided with an optical member 106 called phase shifter. Note, however, that optical member 106 is formed only on a semitransparent film 103 on a glass substrate 101, where a two-layer structure of phase shifter 106 and semitransparent film 103 is formed. Phase shifter 106 serves to change the phase of transmitted light by 180.degree., and semitransparent film 103 serves to attenuate the intensity of exposure light without completely interrupting the exposure light.
Referring to FIG. 16 at (c), since the above-described phase structure is established between phase shifter 106 and semitransparent film 103, the phase of the electric field over the mask is alternately changed by 180.degree., and the intensity of the electric field of one phase is smaller than the intensity of electric field of the other phase. More specifically, the intensity of light is attenuated such that the phase of light transmitted through phase shifter 106 is changed by 180.degree. and then the light transmitted through semitransparent film 103 as well allows the photoresist to remain at a prescribed thickness after development. Since the phases of light of adjacent exposed areas are reversed from each other, these two kinds of light in opposite phases cancel each other in the overlapping regions.
As a result, as shown in FIG. 16 at (d), the phase is reversed at edges of the exposure pattern, and the optical intensity at the edges of the exposure pattern may be reduced. Thus, the difference in optical intensity between the light transmitted through semitransparent film 103 and the light not transmitted therethrough increases, and the resolution of the pattern may be improved as a result.
In FIG. 17, (a) and (b) are a top view and a cross section showing a Levenson-type phase shift mask, and (c) and (d) show the electric field over the phase shift mask in use and the optical intensity of an optical image on the wafer, respectively. Referring to (a) and (b), when a contact hole pattern is formed in a Levenson-type phase shift mask, a transmission region 110 corresponding to a contact hole and a transmission region 111 surrounding transmission region 110 are formed in the mask. Transmission region 110 is provided with an optical member 105 called phase shifter.
On a glass substrate 101, a chromium mask pattern 103 is formed, on which transmission regions 110 and 111 and a light shielding region are formed, with transmission region 111 being provided with phase shifter 105. Phase shifter 105 serves to change the phase of transmitted light by 180.degree..
Referring to FIG. 17 at (c), phase shifter 105 provided in transmission region 111 as described above allows the electric field on the mask created by exposure light transmitted through the phase shift mask to have its phase alternately inverted by 180.degree.. Exposure light is opposite in phase between transmission region 110 and transmission region 111, and therefore, the exposure light cancel each other in the overlapping portion by the effect of optical interference.
As a result, as shown in FIG. 17 at (d), the intensity of the exposure light decreases between transmission region 110 and transmission region 111, and therefore, adequate difference in the intensity of the exposure light between the exposed area and the unexposed area on the wafer may be secured. Thus, the resolution may be improved and fine patterns may be transferred.
However, in order to form a contact hole pattern with a Levenson-type phase shift mask as described above, as shown in FIG. 17 at (a), transmission region 111 having phase shifter 105 surrounding transmission region 110 is necessary in addition to transmission region 110 corresponding to the contact hole. This precludes fine and densely located contact holes from being formed, and the position of a contact hole pattern is limited.
Phase shifter 105 should be provided so as to have a specified size in a specified area and with no defects such as local chips. More complicated mask patterns, however, make checking of such defects more difficult, resulting in increased time and cost for manufacturing phase shift masks.
On the other hand, the Levenson-type phase shift mask has the following advantages.
Referring to FIG. 18 at (a) and (b), the phase shift mask is provided with a line-space pattern. A plurality of light shielding films 103 having a fixed width extend in one direction, and transmission regions 112 between light shielding film 103 are provided with a phase shifter 105 for every other region.
As for the electric field of exposure light transmitted through such a Levenson-type phase shift mask, as shown in FIG. 18 at (c), electric fields directed in opposite directions are alternately distributed. Thus, the intensity of the exposure light on the wafer attains a distribution as shown in FIG. 18 at (d), a distribution having a high contrast. More specifically, by the Levenson-type phase shift mask, adequate contrast is advantageously secured for such a fine line-space pattern.
Reference 1 (Japanese Patent Laying-Open No. 4-22954) discloses a method of forming a fine contact hole pattern with a usual photomask. The method will now be described in detail.
According to the patterning method, an interference aligner is used as an exposure device. FIG. 19 is a view schematically showing the configuration of the aligner. Referring to FIG. 19, the aligner includes a light source 201 emitting exposure light, an expander 202 for expanding the light emitted from light source 201, mirrors 203 and 206 bending the optical path of the exposure light, a half mirror 204 for reflecting part of the exposure light, a phase shift portion 205 for changing the phase of the exposure light, lenses 210 and 211 for paralleling the exposure light transmitted through mask 209, a half mirror 213 for interfering and combining the exposure light transmitted through mask 209, and a reduction lens 214 for reducing the exposure light resulting from the interference.
Now, the method of patterning using the aligner and photomasks will be described.
The exposure light emitted from light source 201 is separated into two beams of exposure light by half mirror 204. One beam of exposure light is directed upon region 209a in photomask 209. The other beam of exposure light is transmitted through phase shift portion 205 to have a phase difference of 180.degree., for example, from that one beam of exposure light, and directed to region 209b in photomask 209 by mirror 206.
Regions 209a and 209b in photomask 209 are provided with first and second patterns for example as shown in FIG. 20 at (a) and FIG. 21 at (a), respectively. FIG. 20(b) and FIG. 21(b) are cross sectional views taken along lines 20b--20b and 21b--21b in (a) in FIG. 20 and (a) in FIG. 21, respectively. Referring to FIG. 20 at (a) and (b) and FIG. 21 at (a) and (b), the photomask having a metal mask pattern 103 is formed on glass substrate 101.
The amplitude of exposure light immediately after being transmitted through the first pattern (the electric field over the photomask) has a distribution as shown in FIG. 20 at (c). The amplitude of the exposure light immediately after being transmitted through the second pattern (the electric field over the photomask) has a distribution as shown in FIG. 21 at (c), because it has a phase difference of 180.degree. with respect to the exposure light transmitted through the first pattern.
The two kinds of exposure light transmitted through the first pattern and second pattern interfere and are combined by half mirror 213. The amplitude of thus combined exposure light has a distribution as shown in FIG. 22 at (a). The resulting combined exposure light passes through demagnification lens 214 and is directed upon wafer 216 on stage 215.
The amplitude of the exposure light on wafer 216 has a distribution as shown in FIG. 22 at (b), and the intensity of the exposure light has a distribution as shown in FIG. 22 at (c). Referring to FIG. 22 at (c), the two kinds of exposure light interfere with each other and the intensities of the two kinds of exposure light cancel each other in the overlapping portions of these antiphase exposure light beams. Thus, difference in the intensity of exposure light beams between the exposed region and the unexposed region on the wafer may be secured, and fine patterns may be formed.
Note that a similar method is disclosed by Japanese Patent Laying-Open No. 3-270213 (Reference 2).
Meanwhile, the inventor has proposed a method of forming a fine contact hole by applying the advantages associated with the Levenson-type phase shift mask described above in Japanese Patent Application No. 08-181300 filed on Jun. 20, 1996 entitled "Patterning Method Using Phase Shift Mask". The method will be now described in detail.
In the patterning method, a double exposure is employed in which two phase shift masks are exposed to light.
FIGS. 23 and 24 are plane views schematically showing a first Levenson-type phase shift mask for use in the first exposure and a second Levenson-type phase shift mask for use in the second exposure, respectively. FIG. 25 is a schematic cross sectional view taken along line 25--25 shown in FIG. 23.
Referring to FIG. 23, first phase shift mask 10A for use in the first exposure in the double exposure method is a Levenson-type phase shift mask, for example, having a transparent substrate 1 of quartz and a light shielding film 3 of chromium.
First and second light transmission portions Tn and Ta transmit exposure light in phases 180.degree. different from each other and are disposed alternately with light shielding film 3 inbetween. First and second light transmission portions Tn and Ta and light shielding portion S having light shielding film 3 each have a linear shape substantially parallel to the Y direction in each plan view. The line widths Wn and Wa of first and second light transmission portions Tn and Ta are approximately identical, and the line width Ws of light shielding portion S is finite.
Referring to FIG. 24, second phase shift mask 10B for use in the second exposure in the double exposure has a structure approximately identical to first phase shift mask 10A as described above. Therefore, the cross section taken along 25--25 for second phase shift mask 10B corresponds to the cross sectional structure shown in FIG. 25.
Note, however, that in second phase shift mask 10B, first and second transmission portions Tn and Ta and light shielding portion S are formed in a linear shape substantially parallel to the X direction in the figure. Now, the patterning method using the phase shift mask will be described.
FIG. 26 is a view for use in illustration of the patterning method using the phase shift mask. In FIG. 26, (a) is a schematic plan view showing how the first and second phase shift masks used in the double exposure overlap each other, (b) shows the intensity of light on the wafer surface along lines 26b--26b in FIG. 26 at (a), and (c) shows the intensity of light on the wafer surface along line 26c--26c in FIG. 26 at (a).
Referring to FIG. 26 at (a), a prescribed region of a semiconductor substrate coated with negative type resist is exposed with light through first phase shift mask (FIG. 23) by means of a demagnification projection exposure method, and the image of the mask is transferred thereon in the process of forming a semiconductor integrated circuit. Then, the same region is exposed with light and the image of the second phase shift mask 10B (FIG. 24) is transferred thereon. At the time, the pattern of the second phase shift mask is exposed with light approximately orthogonally to the pattern of the first phase shift mask.
Referring to FIG. 26 at (b), in the overlapping regions of transmission portions Tn or Ta of first and second phase shift masks 10A and 10B, the intensity of light on the wafer is higher than that in the overlapping regions of transmission portions Tn and shielding portions S or Ta and shielding portion S of first and second shift masks 10A and 10B.
Referring to FIG. 26 at (c), the overlapping regions of light shielding portions S of first and second phase shift masks 10A and 10B, the intensity of light on the wafer is lower than that in the overlapping regions of transmission portions Tn or Ta and light shielding portions S of first and second phase shift mask 10A and 10B. The light intensity on the wafer in the overlapping region of light shielding portions S is 0 or close to 0.
As in the forgoing, the two-dimensional distribution of the light intensity in the wafer surface subjected to the double exposure is as shown in FIG. 27.
Referring to FIG. 27, the densely hatched portions represent regions with lower light intensity, while the blank portions represent regions with higher light intensity. 26b--26b, 26c--26c and 26b--26b in FIG. 27 correspond to 26b--26b, 26c--26c and 26b--26b in FIG. 26 at (a), respectively.
After such exposure process, the usual developing process removes the photoresist in the dark portion (the densely hatched region in FIG. 27) for removal in the case of negative type resist, resulting in a hole pattern.
Reference 1 above describing the technique proposed by the inventors still encounters the following problem.
Reference 1 discloses a method of forming a combined pattern of the first pattern as shown in FIG. 20 at (a) and the second pattern as shown in FIG. 21 at (a) as a desired pattern. However, in the intensity distribution of exposure light on the wafer shown in FIG. 22 at (c), if the intensity of exposure light corresponding to the second pattern region is lower than the intensity as a reference intensity for whether or not the photoresist can be removed by development, only the first pattern may be formed as a contact hole pattern with a high contrast. More specifically, whether a combined pattern is formed or a contact hole pattern is formed, it is essential to form the second pattern which permits the transmitted exposure light to have an opposite phase to the exposure light transmitted through the first pattern around the first pattern in order to form a pattern with high contrast.
This gives rise to the same problem associated with the Levenson-type phase shift mask having a contact hole pattern as described in conjunction with FIG. 17, and a fine hole pattern with densely located holes (hereinafter referred to as dense contact hole pattern) cannot be formed or the position for providing the contact hole pattern is limited.
Now, according to the technique proposed by the inventor, negative type photoresist is used in order to form a contact hole pattern. Referring to FIG. 27, in the case of a negative photoresist, the densely hatched portion in the figure, in other words the photoresist in the regions where the intensity of exposure light is almost 0 are dissolved during development, resulting in a contact hole pattern.
The exposure light generally has its intensity attenuated while progressing through resists. Therefore, along line F--F shown in FIG. 27, for example, light intensity SL as a reference intensity for whether or not resist is removed by development, attains a distribution as shown in FIG. 27 with respect to the depth of resist. Therefore, the hole pattern of developed photoresist sometimes takes a reversely tapered shape.
If the exposure light causes halation by the underlying film, the light is sometimes directed to the region which should not be exposed with light. As a result, part of the photoresist which should be dissolved and removed sometimes remains there. As a result, sometimes a desired dense hole pattern cannot be formed.
Generally, since the negative type photoresist allows polymers to swell in the process of development, the positive type photoresist is used for a pattern which requires higher resolution.
In the use of positive type photoresist, however, a hole pattern cannot be formed, as will be described below. The light intensity along line D--D shown in FIG. 26 at (a) has adequate contrast with respect to light intensity SL as shown in FIG. 26 at (c). The intensity of exposure light along lines 26b--26b shown in FIG. 26 at (a) is higher than light intensity SL as shown in FIG. 26 at (b). In other words, adequate contrast cannot be created, and therefore photoresist positioned near lines 26b--26b is dissolved at the time of development. The photoresist at the densely hatched portion shown in FIG. 27 remains, resulting in a dotted pattern.